文献类型：期刊文献

英文题名：Improved Stability Analysis for Delayed Neural Networks

作者：Li, Zhichen[1];Bai, Yan[2];Huang, Congzhi[2];Yan, Huaicheng[3,4];Mu, Shicai[2]

机构：[1]East China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai 200237, Peoples R China;[2]North China Elect Power Univ, Sch Control & Comp Engn, Beijing 102206, Peoples R China;[3]East China Univ Sci & Technol, Key Lab Adv Control & Optimizat Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[4]Harbin Inst Technol, State Key Lab Robot & Syst, Harbin 150080, Heilongjiang, Peoples R China

年份：2018

卷号：29

期号：9

起止页码：4535

外文期刊名：IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS

收录：;EI(收录号：20174604408580);WOS:【SCI-EXPANDED(收录号:WOS:000443083700053)】；

基金：This work was supported in part by the National Natural Science Foundation of China under Grant 61304041, Grant 61272064, and Grant 61673178, in part by the Fundamental Research Funds for the Central Universities under Grant 2016ZZD03, in part by the Shanghai International Science and Technology Cooperation Project under Grant 15220710700, in part by the Shanghai Shuguang Project under Grant 16SG28, in part by the Shanghai Natural Science Foundation under Grant 17ZR1444700, in part by the State Key Laboratory of Robotics and System (HIT) under Grant SKLRS-2016-KF-03, and in part by the Programme of Introducing Talents of Discipline to Universities (the 111 Project) under Grant B17017.

语种：英文

外文关键词：Delayed neural networks (DNNs); new integral inequalities; stability; triple integral terms

摘要：In this brief, by constructing an augmented Lyapunov-Krasovskii functional in a triple integral form, the stability analysis of delayed neural networks is investigated. In order to exploit more accurate bounds for the derivatives of triple integrals, new double integral inequalities are developed, which include some recently introduced estimation techniques as special cases. The information on the activation function is taken into full consideration. Taking advantages of the proposed inequalities, the stability criteria with less conservatism are derived. The improvement of the obtained approaches is verified by numerical examples.